Monday, August 11, 2014

The Fallacy of Composition and the Monetary System

I was tempted to be snarky about this, but that wouldn’t be fair, on laymen or anybody else.

It’s not easy thinking through economic problems, and monetary problems in particular. Slip ups are common, even among seasoned economists. It doesn’t help that standard texts on money and banking are badly wrong on how banks actually operate and how money is created. It’s no wonder then that people have a hard time figuring out what’s going on, and how changes in policy and customer preferences affect the monetary system.

Example 1 comes from a couple of weeks ago (excerpt):

Money in the modern world

THE Bank of England (BoE) has recently published two articles in its quarterly bulletin on “Money in the Modern World” and “How Money is Created” (

The first should be compulsory reading in schools because it is simple and clear explanation of what money is all about. The second is a more technical explanation of how money is created – surprise, surprise – the central bank is not fully responsible!

The first part of this article is absolutely spot on – it’s one of the clearest expositions of modern money creation I’ve ever come across. Tan Sri Andrew gets it – but then proceeds to show that he doesn’t, not really (excerpt):

…In other words, QE leaves banks with both a new IOU from the central bank but also a new, equally sized IOU to consumers (in this case, the pension fund), and the interest rates on both of these depend on bank rate.”

The second misconception was “Why the extra reserves are not multiplied up into new loans and broad money?” This is because the banks are the transmission mechanism (new code word for villians) of QE. The pension fund now holds bank deposits instead of government bonds (now held by the central bank). The commercial banks have a corresponding increase in reserves with the central bank….

…Can someone explain to me this plain fact – would property prices or stock and bond prices be where they are, if the central banks didn’t engage in QE and doubled or trebled their balance sheet since 2007?

The excess liquidity in the whole world would not have happened without the unconventional monetary policy of the advanced central banks.

Independent central banks would have to do better than writing articles for schoolchildren in absolving themselves from the post-crisis asset bubbles.

The pivotal point to think about here is this – once the central bank purchases government bonds from private agents, what do they do with the (newly created) money they receive? Most large bond holders are institutional, not households, and bonds are held as part of an investment portfolio, so using the proceeds for consumption of goods and services is not really in the picture. That’s why quantitative easing generally does not lead to an acceleration of consumer inflation. Here’s where QE is different from money printing (purchase of securities direct from the government) – governments have a nasty habit of spending any money they receive. But what happens with QE instead?

If you’re a pension fund, as in the example used by the BOE, then the answer is obvious – you invest it again, and likely also in government securities. But this just transfers the money to somebody else, because every asset transaction has a buyer and seller. A purchase (investing) of a security involves somebody simultaneously selling (divesting). That somebody else will also then have excess cash, and would then want to invest in something else.

So what we get in effect is a successive bidding up of government bond prices, up to the point where the yields fall below what investors are comfortable with relative to cash, and the round robin of investing/divesting slows and stops. Since government bond prices underpin prices across the whole capital market, the impact overflows into prices of other securities as well.

What we have is something like the old Keynesian fiscal multiplier, applied to asset prices instead of real economic activity. The bottom line here is that the whole point of QE is to raise asset prices. If you want to think about it in really simple terms, printing money (direct central bank purchases of bonds from the government) causes goods and services inflation, while quantitative easing (central bank purchases of bonds from the private sector) causes asset price inflation.

QE induced asset price inflation causes borrowing costs to fall – higher government bond prices cause higher private debt prices AKA reduced yield AKA reduced private borrowing costs. This is a way for a central bank to reduce the overall cost of credit in an economy, even if the short term policy rate is at or near zero. It’s attacking the long end of the yield curve, which would still have a relatively high positive rate even if the short end is flat and zero. QE is a way of bypassing the (broken) transmission of rate changes from short to long.

That’s why Tan Sri Andrew’s comments at the end of the article are rather strange; it’s a bit like criticising a lion for having a mane, or a leopard for having spots. It’s the nature of the beast, not a side effect.

Another very important point: the impact of expansive global liquidity on domestic asset prices depends crucially on the monetary regime a country has. Those countries where bond and property prices have jumped the most, like Hong Kong and Singapore, have monetary regimes explicitly (HK) or implicitly (SG) anchored on US monetary policy. If you’re going to let Bernanke and Yellen run the kitchen in your house, don’t complain when they heat it up (Ok, I couldn’t resist at least one snarky comment).

Example 2 comes from last weekend (excerpt):

Malaysian banks race to get deposits

PETALING JAYA: Malaysian banks are racing to increase their stash of deposits amid concerns on rising loan-to-deposit (LD) ratios.

This comes after CIMB Group Bhd, the second largest lender, fixed its fixed deposit rates at 3.4% for a 12-month tenure two weeks ago from 3.15% previously….

…The analyst said that both Malayan Banking Bhd (Maybank) and Public Bank Bhd may be forced to offer better FD rates to attract deposits, in order to meet their respective loan growth targets of 10% to 12%….

…The race for fixed deposits, which is a source of cheap funds for banks, is due to the rising loans to deposit ratio (LD ratio) in the past five years. A rising LD ratio means that loans are growing faster than fixed deposits….

…CIMB Research said that banks that were not willing to compete on deposit rates would have to settle for slower loan growth due to the inability to secure adequate funding for that….

…The slower expansion of deposits could be due to alternative investment options, continued strong interest in the property market, and low level of interest rates, which could deter consumers from leaving their money in the banks as it is a less attractive option.

“At the moment, the property market is seen as more attractive to invest in, especially for retail investors.

“They rather invest in property rather than in the bank or in the stock market,” an analyst said.

Ok, some quick comments first:

  1. The LD ratio is the ratio of loans to customer deposits (CASA and term deposits), not fixed/term deposits alone;
  2. If you’ve read Tan Sri Andrew’s article, you’ll recognise that loans create deposits, not the other way around. So the idea that banks are going after customer deposits to “fund” their loan growth is wrong;
  3. FD for banks is not “cheap”. It’s actually on the expensive side, as both CASA and money market rates tend to be lower. The real advantage of term deposits is that it is “locked in” – banks don’t have to meet obligations on term deposits until the term expires, so they can use the cash they gain for other things. That’s not true of almost all the rest of their liabilities (CASA for instance, is payable on demand).

The last point in the article is something I want to expand on – the idea that customers searching for higher yielding assets turning to riskier and alternative investments, are driving down deposit growth. It’s an idea that I’ve heard and been hearing repeated fairly widely, even from analysts and economists in multinational research houses. The germ of the same idea is what drives Tan Sri Andrew’s comments at the end of his article.

There’s a grain of truth here, but it’s verra, verra small. In a nutshell, the criticism is the same as in the first article above.

So let’s think this through. Say a depositor with RM1000 in an FD account thinks the rate is too low – heck, it’s not even enough to compensate for inflation. So he takes the money out, and buys some shares on the stock market. Simple right? Aggregate that across the whole economy, and total deposits have to fall. It’s common sense, right?

Not so fast. Let’s take it one step further.

Who does our putative depositor buy the shares from? Say perhaps somebody who’s sat on those shares for a while, and wants to take profits. But that implies that the original act of investment is matched by an equal and opposite act of divestment. The RM1000 proceeds from the share sale goes into the seller’s current or savings account. So all we have from an increase in risky investment behaviour on the part of our original depositor, is a shift in the ownership of bank liabilities, not a decline in them.

If you look at Malaysian data, that looks pretty clear –growth for term deposits is declining, but not growth in current or savings deposits. A bigger shift would be seen if the sellers are institutional, where they have more options in terms of placing their funds (e.g. NCDs, money markets) i.e. we could see an overall shift out of CASA and term deposits, into say money market instruments, in which case we would have an overall decline in deposit growth, and a corresponding increase in wholesale funding, but not much change for the money supply as a whole.

This has happened most every time the economy expands – there is a shift from less liquid forms of bank liabilities into more liquid forms of bank liabilities, as money velocity increases.

So what if we’re talking about property instead of shares or bonds? The story doesn’t change – there’s a buyer and there’s a seller. Investing in property simply shifts the ownership of deposits from one to the other. It doesn’t matter if you’re purchasing on the secondary market or straight from the developer; the cash doesn’t simply disappear, as implied by the “people buying into risky assets” idea.

It’s a classic case of a fallacy of composition. Just because I buy an asset doesn’t mean the money I used to buy it with vanishes. The fact that I spent RMxxx doesn’t remove that RMxxx from the system – it just belongs to someone else, just as the asset they sold now belongs to me.

At bottom, what we’re looking at here is shifts in risk aversion and liquidity preference change the composition of the money supply, not its growth rate. Basically, it’s the same proposition as the pension fund example used above.

But if my viewpoint is correct, why are banks competing for CASA and terms deposits?

Let’s extend my example a little bit, by assuming the banker for our investor and the banker for our seller are different. That implies that Bank A (our investor’s bank) loses RM1000 in term deposits, while Bank B (the seller’s bank) gains RM1000 in a current account deposit. But Bank A still needs a relatively stable source of liquid cash to meet its other liabilities – it therefore wants to ensure that its deposit rates are attractive enough to “pull back” the deposit it has lost.

That’s why the theory of exogenous money (deposits are needed to create loans) remains so stubbornly alive – the way banks behave in the real world doesn’t really disprove it. But, as always, not understanding the true underlying process can have bad consequences, especially if you’re a policy maker.


  1. Yes, but even in "sheltered" Malaysia, you can't wall yourself away or insulate yourself 100 per cent against US monetary policy.

    You might claim, for instance, that the MYR is rock solid stable against a currency or a basket of currencies, but reality shows otherwise.

    Can you be sure that BNM doesn't care a flying fig about what the Fed is doing or going to do?

    1. @TamanTasek

      Don't be silly, of course you can't fully insulate yourself from changes in global monetary conditions. But that's an order of magnitude different from giving up monetary independence entirely as Hong Kong or Singapore does.

      Which leads to the other point: if the MYR is stable against another currency, especially a major currency, that's when I'd start worrying and digging deeper.

      The two currencies that are least volatile against the Ringgit are the SGD and AUS. In both cases however, this is due to external circumstances. For the former, MAS targets the Ringgit as much as the USD, while the AUS (like the MYR) is a commodity currency, i.e. both tend to fluctuate with changes in global commodity prices

    2. Hisham - why do you say that Hong Kong and Singapore have given up "monetary independence entirely"?

      Granted that the SAR has " pegged" the HKD against the USD, but that was a deliberate choice back in the day.

      Fast forward to today and the RMB is playing an increasingly important role in Hong Kong.

      As for Singapore, the MAS manages the SGD against a basket of currencies, the exact composition of which is a state secret. That's a far cry from giving up "monetary independence entirely".

    3. TamanTasek- There's something in economics that we call a trilemma,

      1. A fixed exchange rate
      2. Free capital movement (absence of capital controls)
      3. An independent monetary policy

      An economy can only choose to have two out of three options listed above, it's impossible to have all three at the same time.

      HK has a fixed ER and free capital movement, therefore it wont' have an independent monetary policy.

      SG on the other hand, implicitly control its currency against a basket of others, and free capital movement. Therefore it also cannot have an independent monetary policy.

      Try google "economics trilemma" or "impossible trinity economics"

    4. @TamanTasek

      What Nazri said :)

      The details are quite interesting and critical to understanding how it works.

      1. Hong Kong does not "peg" the HKD to the .USD. Instead what they have is a currency board mechanism. In effect, Hong Kong cannot have an independent monetary policy and is in monetary terms de facto part of the US.

      2. Singapore operates an exchange rate target based on a path of appreciation (not level) against a basket of currencies, which as you noted is kept secret. However, it's not hard to work out what currencies are in that basket. The exchange rates with the lowest volatility against the SGD are the MYR and the USD. Since the Ringgit is non-convertible internationally, appreciation against the MYR is probably based on adjusting the USD rate. In any case, the mechanism by which the path of the SGD is maintained is effectively the same as if there was a level (pegged) target. Large inflows of FX requires MAS to buy FX and sell SGD. This expands the SGD money supply, depressing domestic interest rates. Large outflows of FX require the opposite, which reduces SGD money supply and raises domestic interest rates.

      Since Fed QE expands USD liquidity (and thus prompts outflows of capital from the US and into other economies), this almost automatically increased SG domestic liquidity. In effect, interest rate policy in SG is determined by the Fed.

      MAS can try, and indeed have tried, to sterilise inflows through mopping up domestic liquidity separately from its FX operations through the issuance of SGS, but this can get expensive and doesn't seem to have made much of an impact. There's a reason why Singapore government debt has ballooned in recent years, to more than double that of Malaysia's.

  2. On another topic - what is your opinion of the views expressed by Prof Anat Admati in her book "The Bankers' New Clothes: What's Wrong With Banking and What to Do About It"?

    Should large banks be required, as Prof Admati advocates, " to raise at least 30 per cent of their funding in the form of equity - about six times more than the current average for the largest American banks"? (New York Times report).

    Would this fly in Malaysia, or would it result in "higher interest rates, less lending and slower economic growth"? (New York Times, ibid).

    1. @anon 5.27

      I haven;t read the book, so I can't comment in detail. Off hand though, it doesn't sound like that bad of an idea, though I'd worry about financial inclusion - sounds like a recipe for increasing wealth inequality. Basel III will take banks part way there in any case.